When explaining a qualitative methodology, many fundamental issues regarding the paradigm by which we pursue science become exposed. Qualitative methodology is often contrasted with and/or a reaction against positivism, the claim of objectivity (realism), reductionism, essentialism, laboratory science, and linear (causal) thinking [see (22)]. To reason and build from its own strengths, however: qualitive research can challenge conventional ideas, inspire new theory, and stimulate minority inclusion and social change (23). As the leading question of this study is comparatively novel, we require an approach that allows an open, deep, and broad exploration. Besides this, we are not sure where we will encounter ideas on the ontology of mental disorder, so a qualitative method may be the best way to approach the research questions. It may help to capture the complexity of the issue better than any other method of research like fixed-choice surveys. In other words, a qualitative method will enable a bottom-up research approach from which new meanings and concepts can be generated. A bottom-up approach, however, does not mean that the stance of the researcher is neutral and without theoretical pre-conceptions. We do not think a point and question from nowhere exists, therefore we explicate it. Our explicit stance derives from relational realism or critical realism, inspired by the thoughts and theories of Latour (24) and Hacking (15). This basically means that we move away from positivism and yet are critical toward pure constructivism. Subject and object are interrelated; scientists (subjects) construct knowledge objects. In the words of Hacking (15), something that is real, is simultaneously constructed. We furthermore hold the idea that psychiatric or psychological science should be built from an explicit ontology. We choose Thematic Analysis as a qualitative research tool as it is a well described and demarcated, yet flexible research tool that has a theoretical freedom and can provide a rich and complex account of data.

A total of 80 patients and clinicians from several health care institutions in the Netherlands were enrolled in the study. Recruitment of the 80 patients and clinicians took place at three regional mental health care providers: De Hoop GGZ, GGZ Momentum and the University Medical Centre of Amsterdam. De Hoop GGZ is a mental Health Care institution with a religious (Christian) basis that has locations throughout the Eastern-central part of the Netherlands. GGZ Momentum is a mental Health Care institution with a humanist basis that has locations throughout the South of the Netherlands. This Health Care institution employs a high number of workers with lived experience. The University Medical Centre of Amsterdam is an academic hospital that is based in the Western part of the Netherlands, with an explicit focus on “evidence-based” care.

We recruited clinicians via the professional networks of AK, LDG, and RL. As participating in this study would entail a time investment for clinicians who already have very full calendars, we chose for this form of convenience sampling. With the use of snowball sampling, we included patients (these were the (ex-)patients of the clinicians we interviewed) and more clinicians and their (ex-)patients. All clinicians were selected based on “purposeful selection”: we included them based on the expectation that they were able to provide in-depth and detailed information about the ontology of mental disorder (25). In addition, we used two criteria for including clinicians in this study: they (i) were seeing patients for diagnosis, treatment, or both and they (ii) were licensed to practice under the Dutch BIG-register. In this way a proper level of training and education could be assumed.

Inclusion of patients was based on their motivation to participate in the study. There was no explicit exclusion criterion, although patients that were in the first stages of their treatment were not actively approached. No participant compensation was provided. Travel expenses were reimbursed if needed, but none of the participants asked for reimbursement (possibly because of the online option that emerged because of the COVID-19 pandemic). The final sample consisted of three groups: patients, clinicians (registered mental health care professionals) and individuals who had been both clinician and user of mental health services.

Over the course of 2 years, interviews were conducted with participants, using a semi-structured interview schedule. We chose a semi-structured format to leave room for new, other, and emerging ideas of participants during the interview. Interviews were audio taped, and typically lasted between 60 and 90 min. Interviews took place in Dutch.

The leading question of this investigation was approached from different angles which led to various interview topics. The semi-structured interview schedule was divided into five main topics: (i) concept of disorder (what exactly is mental disorder?); (ii) treatment (what is treated and how is it (best) treated?); (iii) recovery (what is recovery and what is recovered?); (iv) DSM (does a DSM-diagnosis represent a mental disorder?) and (v) outcome measure (what is the right outcome measure in our field?). A translation of the Interview Schedule can be found in Supplementary Table S1.

Participants were informed and recruited about the research by e-mail. In the case of patient inclusion this was done either after their treatment or at the final stage of their treatment. Those who were interested were given an information letter and provided an opportunity to ask questions about study and procedure. Prior to the start of the interview, participants signed an informed consent form. Audio files were made and carefully locked away for an agreed period of time (typically 2 to 4 weeks). Within this agreed period of time, an anonymized transcript was generated after which the audio file was deleted. Information that could reveal the participant’s identity was removed from the transcripts. All transcriptions of audio files were exported to a protected hard drive at Utrecht University Medical Centre. The study was run after review of and approval by the Medical Ethics Committee and the Institutional Review Boards of the participating health care institutions. Interviews took place at the location of the participant’s choosing (either at the health care provider or, because of the outbreak of the COVID-19 pandemic, online).

The analysis consisted of three main steps. In the first step, the focus was on familiarization with the data. Transcripts were read several times and a mind-map of each interview was constructed by one of the researchers who did not conduct the interview. This was done to grab each participant’s core concept of ontology by reading and seeing the whole of the participant’s story. After this, the researcher who conducted the interview checked the mind-map for accuracy. Consensus over findings and interpretations of the mind-maps was reached by discussion.

The second step involved a more thorough analysis of the transcripts using inductive thematic analysis (26). The Qualitative software analysis program MAXQDA was used to assist with the analyses. Initial codes were generated after which transcripts were analyzed a second time, noting emerging and overlapping themes (repeated patterns). We primarily adopted a semantic approach in which the themes were identified within the explicit meanings of the data (26). Transcripts and emerging themes were reviewed by two additional researchers in order to provide external validation of the analysis. Both systematically checked all codes. This ensured that the themes were represented in the source material. In case of discrepancies, consensus was reached by discussion. Main themes and subthemes were identified, reviewed, refined, and ultimately defined. Although more instances of a theme across one interview and across the whole dataset does not necessarily mean the theme is more substantial, the focus was on themes with a certain weight and frequency across the entire data set (frequency > 10 times). This was done to make sure to not capture idiosyncratic conceptualizations but capturing ideas that live more broadly in the field. These steps resulted in the identification of various themes and an overview of the frequencies of these themes across the whole of the dataset as well as per participant group. In this step, we also decided to omit the topic of “representation by DSM.” Although that the topic is considered highly important, we discovered that the topic was too broad and encompassing too many areas, defying meaningful synthesis for the purpose of answering the leading question of the study.

In the third step, we looked for patterns in the themes across all topics that we identified in the previous step. We looked for the overarching ontological conceptualizations that would answer our leading question. This led to the “identification-construction” of six overarching ontological domains that will be presented in a typology in the results section. We finally checked these over-arching ontological domains with all mind-maps, in order to examine if the typology captured and included the core concept of ontology of participants. Finally, the embedded findings and interpretations were discussed by all authors.

Between September 2018, and December 2020 a total of 80 participants were enrolled in the study who all consented to participate [registered clinicians N = 37 (46%); patients N = 22 (28%); clinicians with lived experience N = 21 (26%)]. Among the registered clinicians 19 (51%) were women and 18 (49%) were men. The mean age of the clinician group was 43.9 (SD 11.4) years. Among the group of registered clinicians 12 (32%) were psychiatrists and 27 (68%) were psychologists. The ratio of psychiatrists to psychologists is about 1:5 in the Netherlands. Of the patient group, 12 (55%) were women and 10 (45%) were men. The mean age of the patient group was 39.7 (SD 11.7) years. Among the clinicians with lived experience of mental health treatment 16 (76%) were women and 5 (24%) were men. The mean age of the clinicians with lived experience of mental health treatment was 34.4 (SD 11.6) years. All three groups had a higher proportion of women than men. During the interview it transpired that several clinicians received treatment for mental distress in the form of psychotherapy/pharmacotherapy. These individuals were placed in the “clinician with lived experience” group. We did not ask about the nature of their condition but let them choose to share this information or not. Clinicians were not placed in the lived experience group if they had sought other forms of help such as coaching, meditation retreats, etc. Table 1 displays the participants’ characteristics.

The inductive thematic analysis of the 80 interviews resulted in 52 subthemes (bottom-up data patterns), 23 main themes (patterns in subthemes) and 6 overarching domains (pattern over all main themes). The domains will be presented in the next section, the main themes per topic are shown in Table 2 below and all 52 subthemes can be found in Supplementary Table S2. Ontologies were most explicitly expressed by participants when asked about conceptual issues (“what is mental disorder?”), however their ontology was also expressed when discussing other topics. Ontologies were often stabile across the interview: interviewees typically held a concept of mental disorder that was also present in their ideas on treatment, recovery, and outcome. If, for example, a participant held that mental disorder is a highly subjective phenomenon, he/she would often also express that he/she would treat a subjective and personalized goal, that recovery is a subjective process and that the right outcome measure should be subjective and personalized in nature. Participants typically held two or sometimes three ontologies that co-existed harmoniously together (e.g., mental disorder as a loss of adaptation and as functional impairment).

This was, however, not always the case. Sometimes seemingly inconsistent ontologies were identified within one interview. For example, one interviewee held an ontological concept of a “diseased brain,” but defined recovery as “a process of acceptance, growth, and gaining autonomy.” At the group level, other interesting inconsistencies were present. As shown in the participant frequencies per topic (Table 2), a substantial proportion of participants held a concept of “deviation from social norms,” yet this fundamental idea on what mental disorder is did not translate into ideas about treatment, recovery, and outcome. Another interesting observation from participant frequencies per topic is that recovery was defined quite broadly and inclusively compared to the answers on other topics. Thus, participants named many things when asked about what recovery is (e.g., reduction of symptoms as well as being able to function as well as finding one’s narrative as well as a highly personal trajectory). Thus, at individual as well as group level, ideas at the level of concept most often resonated with ideas about treatment, recovery and outcome or the other way round, but this was not always the case. With these data, we cannot tell whether practice translated into concepts or the other way round, but we observed consistency in most instances.

The first kind of ontological conceptualization that was found, is well known and can be best described as a “medical-disease conceptualization.” Mental disorder is understood as a thing that typically resides and/or originates in the brain of the individual. In that sense it is an objectivist, realist account in which mental disorders are understood as natural kinds that are real entities that can be objectively observed. Language that typically accompanies this kind of ontological conceptualization includes words like “illness,” “brain-disease,” “biological substrate,” “symptoms,” “response,” “remission,” “symptom reduction,” “symptomatic remission,” and “clinical threshold.” It is often described as something that is determined and static in spatial and temporal terms. Interestingly, in this ontological stance some interviewees differentiate between psychiatric disorders and psychological disorders in which psychiatric disorders are understood as illnesses of the brain and psychological disorders are understood as problems of living [see correspondence to the findings of Ahn et al. (21)]. Within participants, this ontological conceptualization of mental disorder often co-existed with an ontological conceptualization of functional impairment. Here, the criterion for “mental disease” lies in the impairment of functionality or the underlying mental disease can be measured by functional-behavioral impediment. The focus for treatment and recovery in this ontology is remission of the illness/symptoms. Clinicians (psychiatrists) from the university medical center formed the heart and majority of this group. Although there were not many participants that understood the ontology of mental disorder to be a disease, there were many participants that understood recovery as “learning to live with” or self-management of symptoms.

The second kind of ontological conceptualization of mental disorder that was found, can be described as a “functional” ontology of mental disorder. The principle of usefulness or functionality is applied to signify the nature of mental disorder. This ontological stance can be understood as realist and pragmatist as mental disorder is viewed as a problem of functional-practical nature that is objectifiable, real and can be measured. Some participants stressed social-societal or professional roles as most important. Others stressed relational functioning as key, whereas other interviewees underlined emotional or mental functioning as most important. Most often, participants spoke of “general functioning” in which they combined several of these aspects. The beginning and endpoint of this ontological conceptualization was often understood in behavioral terms. The concept and outcome were often defined as (not) being able to “fulfill one’s roles,” “have a work life and social life,” “do what you need or want to do” etc. The larger part of the participants from each of the groups held some sort of “functional” ontology.

The third kind of ontological conceptualization of mental disorder that was found, can be described as an ecological or adaptational ontology. Adaptation refers to the adjustment of humans to become more suited to their environment. In a situation of imbalance by burden, adversity or vulnerability, humans may successfully adapt by finding resources of support and strengths, and by building skills to regain equilibrium and resilience. Examples that were given are improving regulation skills (i.e., self-regulation, emotion-regulation) or finding a suitable and supportive partner, work and friends. People are thought to be able to work on either front; they can improve strengths, recourses and skills or they can lessen their burdens. This ontological conceptualization explicitly incorporates context and fit to context and is systemic in that sense. In this type of ontology, mental disorder is thus conceptualized as not finding or regaining a healthy equilibrium: an imbalance between resources and vulnerabilities remains. This ontological stance can be understood as realist and naturalist as mental (dis)order is conceptualized as a dynamical (eco)system that follows the logic of nature. Although complex, mental disorder is thought to be objectifiable, real and measurable. Imbalance of equilibrium is the point of beginning (concept) and regaining equilibrium is the endpoint (outcome) of this ontological conceptualization. The ecological or adaptational ontology was held by almost half of the sample participants with lived experience and by around a third of the sample clinicians.

The fifth kind of ontological conceptualization of mental disorder that was found, can be described as a “contextual deviation” ontology. This ontological stance can be contrasted with the disease conceptualization in the sense that it (i) places the problem not within an individual but in (between) the context (and an individual), (ii) understands mental disorder not as an illness of the brain but as a deviation from and product of sociocultural norms. This ontological understanding holds that within the small, non-inclusive norm within society, there is no place for variation or deviation. Mental disorder is a product of societal norms; the non-inclusive society creates further deviation mediated by exclusion and loneliness. This ontological stance can be understood as social constructivist and relativist as mental disorder is conceptualized as a “variation” that is produced by society itself. Socio-cultural norms are the point of beginning (concept), yet only very few interviewees (n = 2) mentioned a change of our sociocultural norms as the right endpoint (outcome). Many times, though, the solution to the problem was thought to be within the individual although the problem itself was thought to lie in society. Although around a third of each of the groups held a core concept of social deviation, it was hardly ever the answer they provided when asked what was treated, recovered or what they viewed as the right outcome measure.

The fourth kind of ontological conceptualization of mental disorder that was found can be described as an “existential” ontology. This ontological stance incorporates the concepts of autonomy, narrative, self-governance, self-reliance, sense- and meaning making and freedom of choice. This ontological stance encompasses the idea that mental disorder is about not being able to govern your own life and live according to your own values. In this ontological conceptualization, mental disorder is an experience of loss of autonomy, choice, and meaning over existence. This stance encompasses the idea that the narrative you hold about yourself and live are toxic in the case of mental disorder. In that sense it is a subjectivist, constructivist account in which mental disorders are understood as human constructs that reside in the mind of individuals. In this ontological conceptualization, the point of beginning may be best defined as existential disconnectedness (concept) and the endpoint (outcome) is growth by acceptance and change of meaning and narrative. About a third of patients and clinicians held this ontological stance and about a fifth of clinicians with lived experience held mental disorder to be an existential problem. Although it may be expected that participants from the Mental Health Care institution with a religious (Christian) basis largely formed this ontological stance, this was not the case. Participants that understood mental disorder to be existential in nature were evenly distributed over the three included mental health care institutions.

The sixth kind of ontological conceptualization of mental disorder that was found, can be described as a “subjective-phenomenological” ontology. This stance is non-essentialist at hearth for it holds that mental disorder is a highly subjective phenomenon that differs from disorder to disorder and from person to person in its experience, shape, color, etc. (qualia). In that sense it is a subjectivist, constructivist account in which mental disorders are understood as highly personal phenomena that reside in the minds of individuals. In this ontological stance quantification and objectification of the phenomenon is questioned, if not thought impossible. The subjective feeling of suffering is the point of beginning of this ontological conceptualization and the endpoint (outcome) is highly subjective and personalized in this stance. About a third of patients and clinicians with lived experience held this ontological stance and about two thirds of clinicians held this ontological stance.

In order to give an impression of the differences between sample subgroups in their ontologies, we created a visualization (Figure 2) of subgroup frequencies per ontological domain. Common ground for the three groups in the sample, in terms of ontology, is that mental disorder is about functional impairment. This means that many of the participants from our sample understand mental disorder—at least in part—as functional in nature. However, at the group level, there are some interesting differences. As seen in Figure 2, about a fourth of sample clinicians holds an ontological concept of disease. However, only a small percentage of sample patients and none of the sample clinicians with lived experience adhered to an ontological concept of disease. Also shown in Figure 2, a majority of sample patients and clinicians with lived experience holds an ecological-adaptational ontology, yet a majority of sample clinicians holds a “subjective-phenomenological” ontology. Thus, the clinicians most often understand mental disorder to be a highly subjective phenomenon and individuals with lived experience (patients or clinicians), most often understand mental (dis)order to be about (im)balance of burden and strengths, skills, and recourses. Both groups think that mental disorder is, to some extent, about functional impairment. Within the sample subgroups, many differences in ontological stance are present. This gives rise to the question what this and the above implies for clinical care, research, policy and maybe even education.

Besides these fundamental reflections, several methodological limitations of the current study should be taken into consideration. First, from the pattern in the data, we moved to hypothesis and theory. Therefore, our conclusion drawn from this inductive method cannot be “proven” and remains an educated guess. However, for a qualitative study, we did include quite a large number of participants which makes our findings and conclusions more representative.

The second limitation of this study may be the operationalization. We included clinicians from the hospital and only two secondary mental health care institutions from the Netherlands. In addition to this, we only included patients from secondary mental health care institutions as patients from the hospital were often in the stage of crisis or at their first stages of treatment. One may consider this as an operational limitation as the included participants may not be representative of the ideas that live in their group and in the field; they may be considered to speak from a specific context, organizational culture, or limited experience.

However, as can be seen from the sample characteristics, all clinicians and patients were asked about their past workplaces and past treatments. In these past workplaces and past treatments, most often a very diverse experience is present. Further, the Netherlands may be considered a liberal society and all interviews were anonymous, so we think that participants felt the liberty to think for themselves and from this wide experience. Therefore, this operational aspect likely will not represent a major problem.

We nevertheless think it is important to replicate the current findings in other and more samples, as to test the presented ontological typology. In future research, it may be interesting to look at gender differences and study other and more samples to be able to compare the ontological ideas of different diagnostic samples (e.g., persons with psychotic disorders vs. anxiety disorders). In addition, it would also be interesting to look at moderators like, for example, diagnosis, gender, or ethnical background. Although the findings may not be conclusive, this study does show that unexplored and underserved perspectives on the ontology of mental disorders live among patients and clinicians. Policymakers should invest in the development, elaboration and coming of age of these other ontologies. Educational discourse should at least openly discuss the controversy around the psychiatric object and show and discuss this broader ontological palette. Educational discourse should be more inclusive regarding differing opinions. We need collaborative innovation in the field of psychiatry in which clinicians and patients co-create ideas on the “psychiatric object” that may drive and steer a landscape of promising new approaches. For clinical practice, this may ultimately and hopefully translate into more transparent dialogs and attuned treatment efforts.